{"id":"https://openalex.org/W7171620484","doi":"https://doi.org/10.1145/3794763.3794798","title":"Self-Improving Code Generation via Semantic Entropy and Behavioral Consensus","display_name":"Self-Improving Code Generation via Semantic Entropy and Behavioral Consensus","publication_year":2026,"publication_date":"2026-04-12","ids":{"openalex":"https://openalex.org/W7171620484","doi":"https://doi.org/10.1145/3794763.3794798"},"language":null,"primary_location":{"id":"doi:10.1145/3794763.3794798","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3794763.3794798","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 34th IEEE/ACM International Conference on Program Comprehension","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3794763.3794798","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5143909617","display_name":"Huan Zhang","orcid":"https://orcid.org/0009-0000-0644-3775"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huan Zhang","raw_affiliation_strings":["State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0009-0000-0644-3775","affiliations":[{"raw_affiliation_string":"State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100710505","display_name":"Wei Cheng","orcid":"https://orcid.org/0000-0001-6128-7293"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Cheng","raw_affiliation_strings":["State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0001-6128-7293","affiliations":[{"raw_affiliation_string":"State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100727084","display_name":"Wei Hu","orcid":"https://orcid.org/0000-0003-3635-6335"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Hu","raw_affiliation_strings":["State Key Laboratory for Novel Software Technology, National Institute of Healthcare Data Science, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0003-3635-6335","affiliations":[{"raw_affiliation_string":"State Key Laboratory for Novel Software Technology, National Institute of Healthcare Data Science, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I881766915"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"49","last_page":"60"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":null,"topics":[],"keywords":[{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.5281999707221985},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.49810001254081726},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4740999937057495},{"id":"https://openalex.org/keywords/entropy","display_name":"Entropy (arrow of time)","score":0.46639999747276306},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.4496000111103058},{"id":"https://openalex.org/keywords/code-generation","display_name":"Code generation","score":0.384799987077713},{"id":"https://openalex.org/keywords/test","display_name":"Test (biology)","score":0.3393000066280365},{"id":"https://openalex.org/keywords/program-optimization","display_name":"Program optimization","score":0.3370000123977661}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7505999803543091},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.5281999707221985},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.49810001254081726},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4973999857902527},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.48890000581741333},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4740999937057495},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.46639999747276306},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4496000111103058},{"id":"https://openalex.org/C133162039","wikidata":"https://www.wikidata.org/wiki/Q1061077","display_name":"Code generation","level":3,"score":0.384799987077713},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.3393000066280365},{"id":"https://openalex.org/C139571649","wikidata":"https://www.wikidata.org/wiki/Q1156793","display_name":"Program optimization","level":3,"score":0.3370000123977661},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3303000032901764},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.31380000710487366},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3107999861240387},{"id":"https://openalex.org/C128942645","wikidata":"https://www.wikidata.org/wiki/Q1568346","display_name":"Test case","level":3,"score":0.30059999227523804},{"id":"https://openalex.org/C111065885","wikidata":"https://www.wikidata.org/wiki/Q1189053","display_name":"Fuzz testing","level":3,"score":0.29840001463890076},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.28450000286102295},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.2775000035762787},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C2780522230","wikidata":"https://www.wikidata.org/wiki/Q1140419","display_name":"Ambiguity","level":2,"score":0.26089999079704285}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3794763.3794798","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3794763.3794798","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 34th IEEE/ACM International Conference on Program Comprehension","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3794763.3794798","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3794763.3794798","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 34th IEEE/ACM International Conference on Program Comprehension","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3137102087","display_name":null,"funder_award_id":"62272219","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":31,"referenced_works":["https://openalex.org/W1965194038","https://openalex.org/W1971650562","https://openalex.org/W2041713059","https://openalex.org/W4220699888","https://openalex.org/W4284690374","https://openalex.org/W4311887664","https://openalex.org/W4387321091","https://openalex.org/W4389523706","https://openalex.org/W4399601909","https://openalex.org/W4400582690","https://openalex.org/W4401042444","https://openalex.org/W4401996408","https://openalex.org/W4402671745","https://openalex.org/W4402672073","https://openalex.org/W4402672108","https://openalex.org/W4403536415","https://openalex.org/W4403536418","https://openalex.org/W4405968021","https://openalex.org/W4411113098","https://openalex.org/W4411113383","https://openalex.org/W4411552119","https://openalex.org/W4412886942","https://openalex.org/W4412888419","https://openalex.org/W4412888491","https://openalex.org/W4412944903","https://openalex.org/W4412945242","https://openalex.org/W4412945609","https://openalex.org/W4415797408","https://openalex.org/W7133198585","https://openalex.org/W7133208539","https://openalex.org/W7133243119"],"related_works":[],"abstract_inverted_index":{"Improving":[0],"the":[1,111,122,149,169],"code":[2,67,99,181],"generation":[3,182],"capabilities":[4],"of":[5,19,114,151,171],"large":[6],"language":[7,68],"models":[8,29],"(LLMs)":[9],"typically":[10],"relies":[11],"on":[12,156],"supervised":[13],"fine-tuning":[14,136],"or":[15,30],"preference":[16,131,141],"optimization,":[17],"both":[18],"which":[20],"require":[21],"costly":[22],"external":[23,184],"resources":[24],"such":[25],"as":[26],"powerful":[27],"teacher":[28,77],"reliable":[31],"test":[32,47,53,80],"units.":[33],"However,":[34],"in":[35,179],"real-world":[36],"scenarios,":[37],"it":[38],"is":[39],"much":[40],"harder":[41],"to":[42,74],"obtain":[43],"reference":[44],"solutions":[45],"and":[46,52,78,159,176],"oracles":[48],"than":[49],"problem":[50],"descriptions":[51],"inputs.":[54],"In":[55],"this":[56],"paper,":[57],"we":[58,85,97,127],"tackle":[59],"a":[60,66,75,79,88,102,118,134],"challenging":[61],"yet":[62],"realistic":[63],"question:":[64],"Can":[65],"model":[69],"improve":[70],"itself":[71],"without":[72,183],"access":[73],"superior":[76],"oracle?":[81],"To":[82],"answer":[83],"this,":[84],"propose":[86],"ConSelf,":[87],"self-improving":[89],"approach":[90],"built":[91],"upon":[92],"two":[93],"key":[94],"ideas.":[95],"First,":[96],"introduce":[98],"semantic":[100,172],"entropy,":[101],"novel":[103],"metric":[104],"that":[105,138,163],"measures":[106],"problem-level":[107],"uncertainty":[108],"by":[109,143],"assessing":[110],"functional":[112],"diversity":[113],"program":[115],"behaviors,":[116],"enabling":[117],"curriculum":[119,174],"construction":[120,175],"with":[121],"most":[123],"learnable":[124],"problems.":[125],"Second,":[126],"present":[128],"consensus-driven":[129,177],"direct":[130],"optimization":[132,178],"(Con-DPO),":[133],"preference-based":[135],"method":[137],"weights":[139],"each":[140],"pair":[142],"its":[144],"behavioral":[145],"consensus,":[146],"thereby":[147],"mitigating":[148],"impact":[150],"noisy":[152],"self-generated":[153],"supervision.":[154,185],"Experiments":[155],"various":[157],"benchmarks":[158],"backbone":[160],"LLMs":[161],"demonstrate":[162],"ConSelf":[164],"significantly":[165],"outperforms":[166],"baselines,":[167],"validating":[168],"effectiveness":[170],"entropy-based":[173],"improving":[180]},"counts_by_year":[],"updated_date":"2026-07-31T08:31:51.225901","created_date":"2026-07-30T00:00:00"}